Hi Geert, > > But "mtd-rom" doesn't really have anything to do with the > > functionality of the driver when it is being used in "SPI mode". > > Correct. But DT describes hardware. If the FLASH is used in direct mapped > mode, that should be described in DT. So it seems we are going back to my original plan: Always enable the SPIBSC node in the .dts to ensure the clocks stay on for the XIP kernel case. But, if there is no spi-nor partitions in the node, then simple drop out and do not register the device as a SPI controller. However, you can have mtd-rom partitions in there which is what XIP Linux will use, but we still don't need to register a spi controller for that. Do you agree????? > > > interrupts? RZ/A2M seems to have an SPIBSC interrupt, RZ/A1 hasn't. > > > > There was never any interrupts in the SPIBSC. > > But it looks like when they added HyperFlash and OctaFlash support, > > they put in some interrupts for that. > > And now that I look at it, they are for pins labeled RPC_INT, RPC_WC, > RPC_RESET. > > (I just realized that "RPC" stands for "Reduced Pin Count") > > > > So....am I supposed to add in that interrupt even though I'm not > > planning on using it?? > > DT describes hardware, not driver limitations. OK. I will add the interrupt to the DT. Chris
